Acute effects of ambient air pollution exposure on lung function in the elderly in Hangzhou, China

ABSTRACT

Evidence of an association between acute air pollution exposure and lung function in the elderly is limited. This study is cross-sectional. We quantified the effects of air pollution exposure on lung function among 256 elderly by using a linear mixed model. The results revealed that air pollutants had lag effects on lung function after adjusting for confounders. PM_2.5 (Lag03, Lag 03 was defined three-day moving average, and so forth), PM₁₀, NO₂ (Lag04-Lag05) were significantly associated with reduced FEV₁. PM_2.5 (Lag01-Lag02), PM₁₀ (Lag0-Lag07), NO₂ (Lag0, Lag04), and SO₂ (Lag0) were significantly associated with reduced Forced vital capacity (FVC). PM_2.5 (Lag04-Lag07) and NO₂ (Lag01-Lag07) were significantly associated with reduced FEF_25%–75%. The results showed the adverse change was stronger after adjusting for other pollutants in the PM models, and women were more susceptible to air pollutants. Therefore, we should pay attention to the problem of air pollution in the elderly, especially in women.

KEYWORDS:

• Lung function
• acute effect
• ambient air pollution
• elderly

Introduction

Ambient air pollution is a global challenge that is becoming increasingly common. Substantial evidence shows a strong relationship between exposure to short-term ambient air pollution and the incidence and mortality of the population (Chen et al. 2017; Liu et al. 2019, 2021). Previous research has demonstrated that short-term air pollution exposure has a negative impact on lung function in healthy elderly individuals, leading to decreased lung function (Simoni et al. 2015; Zhang et al. 2019). An IQR increase in PM_2.5 (particulate matter with aerodynamic diameter ≤2.5 µm) was associated with a 73.30 ml decrease in forced expiratory volume in one second (FEV₁) and a 106.38 ml decrease in forced vital capacity (FVC) (Chen et al. 2019). However, most previous studies have focused on children and chronic obstructive pulmonary disease (COPD) patients (Hansel et al. 2016; Xu et al. 2018), and research on the elderly population is still scarce, especially in highly polluted areas. Whether or not the healthy elderly in China have a similar result in short-term air pollution exposure on lung function as children and COPD patients is unknown.

Middle-aged and elderly individuals are generally more susceptible to ambient air pollution exposure than the general population (Chen et al. 2018; Amsalu et al. 2019). Normal aging results in decreased muscle quality and immune system function (Cho and Stout-Delgado 2020). Previous studies have shown that normal lung aging is associated with functional and structural changes in the deep respiratory tract, which leads to decreased lung function and enhanced susceptibility to the lung (Shivshankar et al. 2011; Skloot 2017).

To date, most studies conducted on environmental epidemiology and quantitative methods, such as lung function tests, questionnaire surveys, and ambient air pollution data, have been used to evaluate the association between air pollution exposure and lung function. The lung function test has been widely used in most environmental epidemiology research because of its relatively low price and reliability (Ruppel and Enright 2012), including large and small airway parameters such as FEV₁, FVC, and mid-expiratory flow between 25% and 75% of forced vital capacity (FEF_25%–75%), to quantify the severity of early respiratory injury and physiological state (Miyoshi et al. 2020). Previous studies have reported an association between short-term exposure to ambient air pollution and decreased lung function parameters (Xu et al. 2018; Yang et al. 2020; Zhou et al. 2020).

China is experiencing rapid urbanization and economic growth, air pollutant emissions have increased continuously, and the importance of the environment has also been gradually highlighted (Chen et al. 2017). With the progress of socialization in China, the aging of the population has become increasingly prominent. The 2020 National People Census has reported that 18.70% of the total population are aged 60 years and older, and 13.50% are aged 65 years and older (Tu et al. 2021). As the elderly population increases, we need to understand the health effects of exposure to different ambient air pollutants. Therefore, it is of great importance to promote healthy aging and develop preventive strategies to reduce unnecessary injuries.

This study aimed to explore the association between short-term (up to 7 days) ambient air pollutant exposure and lung function parameters in the elderly.

Methods

Study population

Hangzhou City, which is located in the northern Zhejiang Province of China, belongs to the subtropical monsoon zone, with a total population of approximately 12 million. China National Environmental Monitoring Centre (CNEMC) reported that the annual daily average concentrations of PM_2.5, PM₁₀, and NO₂ in Hangzhou were 42.42 µg/m³, 70.13 µg/m³, and 42.35 µg/m³ between 2015 to 2020, respectively, which were 4 to 6 times higher than those in the 2021 World Health Organization Air Quality Guidelines (PM_2.5 should not exceed 5 µg/m³ annual; PM₁₀, 15 µg/m³; and NO₂, 10 µg/m³).

Lung function tests were conducted between December 2020 and August 2021 in winter and summer, respectively. The interval between these two measurements was approximately six months. Considering the large population of Hangzhou, a two-stage sampling was used in this study. First, four districts were randomly chosen from Hangzhou City on the basis of the probability proportional to their population, and communities with more than 10,000 residents were chosen as basic units. In the second stage, simple random sampling was used to select a community from the district extracted in the first stage, and approximately 60 elderly residents (60–75 years) from each community were randomly selected. For each selected district, we established a table containing the elderly population who met the inclusion and exclusion criteria. The first participant was selected by generating a random number; then, the other elderly population was chosen by equal interval numbers.

The prevalence of COPD was taken as 8.37% (Wang et al. 2018), 95% confidence level, 5% margin of error, 2 design effect, considering a drop-rate of 15%, and finally, 276 subjects were recruited. Ultimately, 265 subjects with a total of 530 lung function data were included in this study.

Inclusion and exclusion criteria

Healthy elderly people of both sexes aged between 60 and 75 years were included in this study. To ensure the reliability of the results, the study population had some limitations. The inclusion criteria were as follows: (a) participants aged between 60 and 75 years, (b) duration of residence in the local community for more than five years, and (c) voluntary participation. The exclusion criteria were as follows: (a) severe mental disease and inability to cooperate with the lung function test, (b) underlying lung disease, such as severe bronchial asthma, bronchiectasis, tuberculosis, and COPD; c) acute lung infectious diseases or acute exacerbations of COPD and a history of antibiotics during the study period; and (d) there were fewer than 1/3 questions answered, and all answered the same.

Questionnaire

We conducted questionnaire surveys for each participant through face-to-face interviews, and the questionnaire was divided into two parts: basic information and disease history. Basic information on the questionnaire included age, sex, height, weight, education level, frequency of cooking at home, residential history, presence of pets and purifiers at home, smoking status, alcohol status, exercise status, and other demographic information. The disease history questionnaire included a history of cardiovascular disease, respiratory disease, and other diseases that may affect lung function.

Lung function test

Lung function tests were performed in local community hospitals, and the test rooms used for testing were quiet and had a suitable temperature. Spirometry was performed by specialists using a Pony FX instrument (COSMED, Rome, Italy) in accordance with the guidelines provided by the American Thoracic Society (ATS)/European Respiratory Society (ERS) (Miller et al. 2005). Each participant was asked to perform the test at least three times during the study, and the best values were selected as the final test result. The Pony FX instrument was calibrated according to the changes in temperature and relative humidity before lung function testing for each participant. Three lung function parameters were selected: FEV₁, FVC, and FEF_25%–75%.

Air pollution and meteorological data

Daily average air pollution concentration data were collected from the nearest fixed-site monitoring stations in the selected area from CNEMC (http://www.cnemc.cn/). In this study, monitoring station data were used as exposure values for the population. One monitoring station was located in each area. All participants recorded in this study resided less than 10 km from the nearest monitoring station. In addition, daily average temperature and relative humidity data were also collected from the China Meteorological Administration to control for the influence of meteorological factors on lung function.

We applied moving-average exposure lag structures to explore the relationship between ambient air pollutants and lung function. Lag 0 day was considered the day of lung function test, lag 01 day was considered the day of lung function test to 1 day before the day of lung function test, lag 02 day was considered the day of lung function test to 2 days before the day of lung function test, and so on.

Statistical analysis

Basic information of participants’ lung function, and air pollutant exposure data are expressed as the mean (standard deviation) or percentiles (P₂₅, Median, P₇₅). Categorical variables were expressed as frequencies and percentages. The chi-square test was used for categorical outcomes, and t-tests were used for continuous variables. Correlations between ambient air pollutants and meteorological data were estimated using the Spearman rank correlation.

The association between acute exposure to ambient air pollutants and lung function was assessed using a linear mixed-effects model. The model included participants and different districts as random effects and was adjusted for gender, age, BMI, season, air pollutants, daily temperature, and relative humidity. Other variables such as educational level, frequency of home cooking, presence of pets and purifiers at home, smoking status, and alcohol status were not included in our model on the basis of the AIC and BIC criteria. The modeling strategy and results were shown in Supplementary Table S1 and S2. The final changes were presented as changes in lung function with a 10 µg/m³ increase in the daily concentration of ambient air pollutants (Wu et al. 2010). The two equations were used as shown below, and β, SE, and ∆x refer to the effect estimate, standard error, and increase in each pollutant, respectively.

Percent change:

(1) $\left(e^{\left(\beta \times \mathrm{\Delta }\chi \right)}-1\right)\times 100\mathrm{\%}$(1)

95% confidence intervals (95% CI):

(2) $\left(e^{\left(\mathrm{\Delta }\chi \times \left(\beta \pm 1.96\times SE\right)\right)}-1\right)\times 100\mathrm{\%}$(2)

Previous studies have shown that air pollutants are correlated, and multi-pollutant models have been used to assess the independent effects of each pollutant. The two-day moving average (lag 01 day) concentration was chosen for the main subgroup analyses, according to previous studies. We performed stratified analyses by sex (male and female) to compare the relationship between lag 01 day air pollutants exposure and lung function.

The following equation was used to determine whether there was a statistical difference between single-pollutant models and co-pollution models: β1 and β2 refer to the effect estimates of the two groups; SE1 and SE2 refer to the standard errors of the effect estimates of the two groups:

(3) $\left({\beta }_1+{\beta }_2\right)\pm \sqrt{S{E}_1^2+S{E}_1^2}$(3)

Statistical significance was defined as P < 0.05 (two-sided). All statistical analyses were performed using R studio version 4.1.0, using the lme4 package.

Results

Basic characteristics of participants

Table 1 shows that out of 265 participants in this study, 11 refused to participate (participation rate: 96%), and the average age of all participants was 68.67 ± 4.36 years with a range from 60 to 75 years and males and females accounting for 41.13% and 58.87%, respectively. The average height, weight, and BMI were 159.01 cm, 59.78 kg, and 23.59 kg/m³, respectively. Forty-eight (18.11%) participants had a smoking history, and 79 (29.81%) had a history of drinking. A total of 67 participants (25.28%) had a history of pets in the household, and 104 (39.25%) had a history of refurbishment. Results are provided in Table 1.

Table 1. Basic characteristics and lung function outcomes of participants.

Characteristics	Total
N (%)	265
Age, years, Mean (SD)	68.67(4.36)
Height, cm, Mean (SD)	159.01(7.58)
Weight, Kg, Mean (SD)	59.78(9.17)
BMI, kg/m^2, Mean (SD)	23.59(2.83)
Gender, N (%)
Males	109(41.13%)
Females	156(58.87%)
Education, years, N (%)
9	192(72.45%)
9 to 12	42(15.85%)
12	31(11.70%)
Smoking status, N (%)
yes	48(18.11%)
no	217(81.89%)
Drinking status, N (%)
yes	79(29.81%)
no	186 (70.19%)
Pet, N (%)
yes	67(25.28%)
no	198(74.72%)
Refurbish house, N (%)
yes	104(39.25%)
no	161(60.75%)
Lung function outcomes, Median (IQR)
FEV1, L	1.89(0.73)
FVC, L	2.43(1.01)
FEF25%–75%, L/s	2.04(1.11)

FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; FEF_25%–75%: forced expiratory volume at 25–75% of forced vital capacity; SD: standard deviation

Descriptive data analysis and correlation analysis

Table 2 shows the results of the daily ambient air pollutants and meteorological data, which presents the mean, standard deviation, minimum, median, and maximum for air pollution and meteorological data. The correlation coefficients between air pollutants and meteorology are also shown in Table 2. The four air pollutants were positively correlated with each other (P < 0.01), and the relative humidity was negatively correlated with the four air pollutants (P < 0.01).

Table 2. Distribution of four air pollutants and meteorological factors, with Spearman’s rank correlation coefficients during the survey.

Variables	Mean（SD）	P25	Median	P75	Spearman’s rank correlation coefficients
					PM2.5	PM10	NO2	SO2	Temp	RH
PM2.5, µg/m^3	32.3(21.7)	18	27	39	-	0.77**	0.59**	0.22**	−0.29**	−0.21**
PM10, µg/m^3	58.6(32.9)	37	52	72		-	0.54**	0.24**	−0.37**	−0.41**
NO2, µg/m^3	32.1(22.5)	16	31	42			-	0.21**	−0.48**	−0.25**
SO2, µg/m^3	13.0(21.4)	4	6	9				-	0.16**	−0.28**
Temp, ℃	19.6(8.9)	11.4	22.0	27.3					-	0.08
RH, %	71.8(14.9)	62.2	73.2	84.5						-

PM_2.5: particulate matter with aerodynamic diameter ≤2.5 µm; PM₁₀: particulate matter with aerodynamic diameter ≤10 µm; NO₂: nitrogen dioxide; SO₂: sulfur dioxide; Temp: temperature; RH: relative humidity.

**P < 0.01; *P < 0.05.

Association between air pollution and lung function change

Figure 1 shows the percentage changes in lung function per 10 µg/m3 increase in a single pollutants with moving day average exposure. Specific data were shown in supplementary Table S3. PM_2.5 (Lag03), PM₁₀ (Lag04-Lag05), and NO₂ (Lag04-Lag05) were significantly associated with a decreased FEV₁. PM_2.5 (Lag01-Lag02), PM₁₀ (Lag0-Lag07), NO₂ (Lag0, Lag04), and SO₂ (Lag0) were significantly associated with a decreased FVC. PM_2.5 (Lag04-Lag07) and NO₂ (Lag01-Lag07) were significantly associated with decreased FEF_25%–75%. For the four air pollutants, the lag trends in FEV₁ and FVC were similar. The damage caused by NO₂ on FEF_25%–75% becomes greater with increasing moving average exposure time. Double interactions between predicting variables were shown in Supplementary Table S4.

Figure 1. Percent changes with 95% confidence interval in lung function parameters associated with 10 µg/m³ increase in a single pollutant with moving day average.

These are estimates from our models, with adjustments for sex, age, BMI, season, educational level, frequency of home cooking, presence of pets and purifiers at home, smoking status, alcohol status, daily temperature, and relative humidity. PM_2.5: particulate matter with aerodynamic diameter ≤2.5 µm; PM₁₀: particulate matter with aerodynamic diameter ≤10 µm; NO₂: nitrogen dioxide; SO₂: sulfur dioxide; FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; FEF_25%–75%: forced expiratory at 25%–75% of forced vital capacity

Table 3 shows that the NO₂ and SO₂ multi-pollutant models were not significantly associated with the decrease in lung function, except for the PM_2.5 model on FEF_25%–75%.

Table 3. Percent changes with 95% confidence interval in lung function parameters associated with 10 µg/m³increase, including in single, multi-pollutant models with lag01 day exposure.

Models	FEV1		FVC		FEF25%–75%
	%	95%CI	%	95%CI	%	95%CI
PM2.5	−0.37	(−1.01, 0.28)	−1.14	(−2.01, −0.27)	−.82	(−2.92, 1.32)
+NO2	−0.29	(−1.02, 0.45)	−1.07	(−2.10, −0.03)	−2.76	(−3.95, −1.55)
+SO2	−0.49	(−1.16, 0.18)	−1.27	(−2.15, −0.38)	−4.59	(−5.63, −3.54)
+NO2+SO2	−0.45	(−1.25, 0.36)	−1.27	(−2.33, −0.21)	−5.82	(−6.94, −4.69)
PM10	−0.23	(−0.52, 0.06)	−0.63	(−1.01, −0.26)	0.55	(−.44, 1.55)
+NO2	−0.21	(−0.52, 0.09)	−0.58	(−.97, −0.19)	0.60	(−0.42, 1.63)
+SO2	−0.27	(−0.58, 0.03)	−0.80	(−1.18, −0.42)	0.65	(−0.37, 1.68)
+NO2+SO2	−0.29	(−0.63, 0.04)	−0.81	(−1.24, −0.69)	0.50	(−0.48, 1.49)

The model included participants and different districts as random effects and was adjusted for sex, age, BMI, season, air pollutants, daily temperature, and relative humidity. PM_2.5: particulate matter with aerodynamic diameter r ≤2.5 µm; PM₁₀: particulate matter with aerodynamic diameter ≤10 µm; NO₂: nitrogen dioxide; SO₂: sulfur dioxide; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; FEF_25%–75%: forced expiratory at 25%–75% of forced vital capacity.

Stratified analysis

Table 4 presents the results of the stratified analyses. The results revealed that PM_2.5, PM₁₀, and NO₂ more strongly affected FEV₁, FVC, and FEF_25%–75% in females than in males.

Table 4. Percent changes with 95% confidence interval in lung function parameters in single pollutant models with lag01 day exposure.

Lung function	Pollutants	Male		Female
		%	95%CI	%	95%CI
FEV1	PM2.5	1.20	(−0.39, 2.82)	−1.94**	(−3.49, −0.37)
	PM10	1.57	(0.97, 2.17)	−2.03**	(−2.80, −1.26)
	NO2	1.67	(0.49, 2.88)	−2.35**	(−3.40, −1.29)
	SO2	0.63	(−0.59, 1.86)	−1.02	(−2.18, 0.15)
FVC	PM2.5	0.60	(−0.84, 2.12)	−3.05**	(−4.47, −1.62)
	PM10	1.84	(0.61, 3.09)	−3.10**	(−4.14, −2.05)
	NO2	2.08	(−1.27, 5.54)	−3.78**	(−4.89, −2.66)
	SO2	−0.33	(−1.46, 0.81)	−1.35	(−2.40, −0.27)
FEF25%–75%	PM2.5	1.36	(−1.58, 4.38)	−3.00**	(−5.63, −.31)
	PM10	1.66	(0.34, 3.01)	−.55**	(−1.90, 0.82)
	NO2	0.21	(−3.48, 4.03)	−6.63**	(−9.90, −3.24)
	SO2	−0.16	(−5.51, 2.59)	−1.54	(−5.14, 2.19)

The model included participants and different districts as random effects and was adjusted for sex, age, BMI, season, air pollutants, daily temperature, and relative humidity. PM_2.5: particulate matter with aerodynamic diameter ≤2.5 µm; PM₁₀: particulate matter with aerodynamic diameter ≤10 µm; NO₂: nitrogen dioxide; SO₂: sulfur dioxide; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; FEF_25%–75%: forced expiratory at 25%–75% of forced vital capacity.

** P < 0.01; *P < 0.05.

Discussion

Our study aimed to determine the association between acute ambient air pollutants (PM_2.5, PM₁₀, SO₂, and NO₂) and lung function parameters. We found that short-term exposure to ambient air pollutants was associated with decreased lung function parameters (FEV₁, FVC, and FEF_25%–75%). For PM_2.5, the effect change was stronger after adjusting for other air pollutants in the FEF_25%–75%. However, the effect weakened with a longer lag time for NO₂. Sex-stratified results showed that women were more susceptible to air pollutants. To our knowledge, this is the first study to examine the association between short-term ambient air pollutant exposure and lung function parameters in an elderly population (60–75 years).

We found that PM_2.5 exposure was significantly associated with a decrease in FEF_25%–75% for up to 7 days, indicating that PM_2.5 has a greater impact on small airway lung function (FEF_25%–75%) than on big airway lung function (FEV₁ and FVC). Evidence suggests that PM_2.5 can produce inflammatory cells and inflammatory mediators, including cytokines, chemokines, and adhesion molecules, which can lead to reduced small airway lung function (Zhang et al. 2009; Kelly and Fussell 2011). A multi-center cohort study and meta-analysis(Adam et al. 2015) found that an increase of 10 µg/m³ in PM₁₀ was associated with a decrease in FEV₁ (−44.6 mL, 95% CI: −85.4, −3.8) and FVC (−59.0 mL, 95% CI: −112.3, −5.6), and PM₁₀ has a greater effect on FVC than FEV₁. Although Adam’s study is a cohort study based on long-term exposure, these results are similar to those of our study: PM₁₀ was associated with decreased FEV₁, and this effect was enhanced with lag time.

In our study, we found a significant association between NO₂ concentration and a decrease in FEF_25%–75%, similar to the results of previous studies (Lagorio et al. 2006; Zhou et al. 2016). Animal experiments have shown that short-term NO₂ exposure can cause lung injury by inflammatory responses and an imbalance in Th1/Th2 differentiation ratios (interleukin-4, interferon gamma, GATA-3, and T-bet) (Ji et al. 2015). Several studies have reported a significant association between SO₂ and decreased lung function (Steinvil et al. 2009; Hong et al. 2018). However, we found that SO₂ was not associated with FEV₁, FVC, or FEF_25%–75%, which may be because the concentration of SO₂ was too low. A population-based study in a Chinese population showed that low-level SO₂ exposure had a protective effect on lung function (Ge et al. 2017). Moreover, similar results were obtained in another animal study; SO₂ played an important role in the process of lipopolysaccharide-induced acute lung injury, suggesting acute protective effects of low-level ambient SO₂ exposure on bacteria-induced lung function infections (Meng et al. 2003; Ma et al. 2012).

We found that the changes in FEF_25%–75% were estimated to be the greatest in the multi-pollutant model. Some studies found that PM_2.5 exposure could exacerbate the damage in lung function after adjustment for SO₂ and NO₂ (Faustini et al. 2012; Chen et al. 2015). FEF_25%–75% is a useful parameter for predicting abnormal lung function, has a high application value, and is widely used in lung function studies (Riley et al. 2015; Kwon et al. 2020).

Previous studies have shown that girls are more susceptible than boys to ambient air pollution exposure (Chen et al. 2018; Yang et al. 2020). In our study, the female population was more susceptible to air pollutants than the male population. Lung function levels in females are significantly lower than males due to differences in lung development and body structure (Paulose-Ram et al. 2015). A study from France, conducted in the elderly, showed that women were more sensitive to PM₁₀ and SO₂ exposure after adjusting for potential confounders (Bentayeb et al. 2010). It is important to note that the results were not restricted to the healthy elderly population but were also present in the general elderly population.

Our study has some limitations. First, we used fixed monitoring results as a proxy for population exposure to air pollution. Although we believe it is reasonable to use these measurements as a good proxy for population exposure, the discrepancy between the results of these fixed monitoring stations and the true exposure values of individuals is an inherent and unavoidable type of measurement error. Second, The study has several limitations regarding the definition of the exclusion criteria, for example, participants with current respiratory tract infections, or patients recovering from acute infections were not included in the exclusion criteria, and may have impacted on the results. Third, this study was conducted in only one city, and future epidemiological studies are needed in different cities, especially in areas with high pollution risk. Finally, this study is similar to across-sectional study and is more prone to bias in the results than along-term cohort study. Future studies should aim to establish long-term cohorts.

Our study has several strengths. First, this study included only a healthy elderly population. Previous similar studies have focused on people with diseases, with relatively few studies on healthy elderly people. Second, we used various lung function parameters to investigate the effects of short-term ambient air pollutants on lung function, whereas most previous studies have adopted a single type of lung function parameter(Lee et al. 2007; Hou et al. 2020; Zhou et al. 2020).

Conclusions

In conclusion, we found that short-term ambient air pollutant exposure was associated with decreased lung function parameters among healthy elderly individuals, where the correlation almost always lagged. Gaseous pollutants not only affect lung function but also enhance the association between PM pollutants and lung function parameters. Sex-stratified results showed that women were more susceptible to air pollutants. These results emphasize that the elderly population, especially females, should be aware of the dangers of air pollutants. Protective measures are required for exposure to ambient air pollutants for one week.

Acknowledgement

We thank all participants involved in the present study.

Disclosure statement

The study was approved by the ethics committee of the Hangzhou Center for Disease Control and Prevention (CDC) (number 2020–3). All participants were informed and written consent before the study began.

HongXu (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported, no important aspects of the study have been omitted; any discrepancies from the study as planned (and, if relevant, registered) have been explained.

The Corresponding Author has the right to on behalf of all authors and does grant on behalf of all authors.

Additional information

Funding

This work was supported by the Key Project of Health and Medicine Program of Hangzhou under Grant No. OO20190549 and No. ZD20200037; Scientific and Technological Projects of Hangzhou under Grant No. 20191203B142.